vulnerability and adaptation of wheat to climate change in ... - IWTC

Thirteenth International Water Technology Conference, IWTC 13 2009, Hurghada, Egypt

VULNERABILITY AND ADAPTATION OF WHEAT TO CLIMATE CHANGE IN MIDDLE EGYPT Fouad A. Khalil(1), Hanan Farag(2), Gamal S. El Afandi(3) and Samiha A. Ouda*(1) (1)

(2) (3)

Water Requirements and Field Irrigation Research Department; Soil, Water, and Environment Research Institute; Agricultural Research Center; Egypt Environment and Climate Research Institute; National water Research Center; Egypt Al Azhar University; Faculty of science; Department of Astronomy and Meteorology; Egypt

ABSTRACT The effect of climate change on the yield of three wheat varieties (Sids1, Sakha 93 and Giza 168) and consumptive use was studied by implementing two-year field experiment in Giza Agricultural Station, Giza, Egypt in 2006/07 and 2007/08 growing seasons using CropSyst model with two climate change scenarios. These scenarios were A2 (temperature increase by 3.1°C and CO2 concentration is 834 ppm) and B2 (temperature increase by 2.2°C and CO2 concentration is 601 ppm) developed by Hadley Center for Climate Prediction and Research. CropSyst model was validated using the collected data of wheat yield and consumptive use. The scenarios were used to run the CropSyst model and to predict the expected yield in the year of 2038. Two early sowing dates were proposed as adaptation options, i.e. 1st of November and 21st of October to reduce the harm effect of climate change on wheat yield and a new irrigation schedule was used. The results indicated that CropSyst predictions for yield and consumptive use were highly accurate. Furthermore, A2 scenario predicted greater reduction in wheat yield, compared with B2 scenario in the year of 2038. Likewise, wheat yield losses were higher at the 1st growing season, compared with the 2nd growing season under the two scenarios. The results also revealed that under the 1st growing season for both climate change scenarios, Sakha 93 variety was found to be more tolerant to heat stress. Whereas, Sids 1 variety was found less vulnerable to climate change in the 2nd growing season. The results also showed that wheat yield improvement and irrigation water saving could be attained using the proposed adaptation strategies in the year of 2038. Under cultivation in November, 1st, a slight improvement in yield losses could be achieved with a slight increase in the amount of applied irrigation water. Whereas, under sowing in October, 21st, a decrease in yield losses could be achieved with a decrease in the amount of applied irrigation water. Under all cases, water use efficiency was increased, compared with its value under the two climate change scenarios.

* corresponding author email: [email protected]


INTRODUCTION Predictions of human induced global climate change are derived from increases in atmospheric level of carbon dioxide. One of its adverse effects is warmer temperatures and increasing episodes of very hot weather. Temperature is the primary factor driving wheat development (Wilhelm and McMaster, 1995), and consequently influence yield (McMaster, 1997). Numbers of tillers are usually decreased when wheat plants were exposed to high temperature (Friend 1965). In addition, temperature is the major variable controlling spikelet initiation and development rates (McMaster, 1997). Furthermore, high temperature during anthesis causes pollen sterility (Saini and Aspinall, 1982) and reduces number of kernels per head, if it prevailed during early spike development (Kolderup, 1979). At higher temperature, the duration of grain filling period was reduced (Sofield et al., 1977) as well as growth rates with a net effect of lower final kernel weight (Bagga and Rawson 1977; McMaster, 1997). Therefore, it is expected that climate change will have implications for possible fluctuation on wheat yield (Wrigley, 2006). Many studies have documented the effects of climate change on agriculture in Egypt and pose a reasonable concern that climate change is a threat to sustainable development. Climate change could do severe damage to agricultural productivity if no adaptation measures are taken (El-Shaer et al. 1997). Most of the previous research on the impact of climate change on agricultural sector used two scenarios, i.e. 1.5°C rise in temperature (MAGICC/SCENGEN results) and 3.6°C rise in temperature (GCM results) to predict the impact at the year 2050. These scenarios predicted reduction in wheat grain yield by up to 30% and increase in its water needs by up to 3% (Eid et al., 1992; Eid et al., 1993 and Eid et al., 1994) in the year of 2050. Thus, the effects of climate change on wheat production will determine the future of food security in Egypt, especially under the existence of large gap between wheat production and consumption. For that reason, adaptation strategies should be explored reduced the vulnerability of the system to climate change. Pervious research suggested that increasing the applied irrigation water amount, increasing nitrogen fertilizers and delay sowing could be used to reduce the vulnerability of crops to climate change (Eid et al., 1995; El-Shaer et al., 1997; Eid and El-Mowelhi, 1998 and Eid and El-Marsafawy 2002). However, warming could also affect water resources and that will pose another problem, which is water scarcity. Furthermore, increasing nitrogen fertilizer could increase the soil and ground water pollution. Whereas, delay sowing could expose the growing plants to higher temperature, which will negatively affecting the final yield. On the contrary, early sowing could help the growing plants to escape heat stress (Wrigley, 2006) and that could result in yield improvement. The objectives of this research were: (i) To use CropSyst model to simulate wheat yield under two climate change scenarios; (ii) To use CropSyst model to test the effect of early sowing as an adaptation option on relieving the harm effect of climate change on wheat yield and water use efficiency.


MATERIALS AND METHODS 1. Field experiments Two field experiments were conducted in 2006/07 and 2007/08 growing seasons in Giza Agricultural Research Station, Egypt to collect data on wheat grain and biological yield. These collected data was used to validate CropSyst model and to run it under two climate change scenarios. CropSyst model was also used in assessing the effect of early sowing and increasing number of irrigations on wheat yield and water use efficiency under the two climate change scenarios. Three wheat varieties were planted, i.e. sids 1, sakha 69 and Giza 128 in a randomize complete plot design with three replicates. Wheat was planed on the 15th and 17th of November in the first and second growing seasons, respectively. Nitrogen fertilizer was divided into 3 doses (at sowing date, tillering stage and at boating stage) in the form of Urea (180 kg/ha, 46% N). Phosphorus fertilizer was applied in the form of single super phosphate (36 kg/ha, 15% P2O5) and was incorporated into the soil during land preparation. Potassium in the form of potassium sulphate (57 kg/ha, 48% K2O) was applied at boating stage. The applied amount of NPK fertilizer was sufficient to ensure optimum growth. Irrigation was applied using 1.2 pan evaporation coefficient, which is the optimum one for wheat under Giza climate conditions. Evaporation data were collected on a daily basis from a standard Class-A-Pan located near the experimental field. Irrigation amounts were calculated with the following equation (Allen et al., 1998): I = Epan*Kp

(1)

Where: I is the applied irrigation water amount (mm), Epan is the cumulative evaporation amount in the period of irrigation interval (mm), Kp is the pan evaporation coefficient. The total number of irrigations was 7 irrigations. Soil mechanical analysis according to Piper, (1950) of the experimental field in the depth of 0-60 cm is shown in Table (1). Table (1): Soil Mechanical analysis at Giza Agricultural Station Soil fraction Coarse sand Fine sand Silt Clay Texture class

Content (%) 2.91 13.40 30.51 53.18 Clay

The soil moisture constants (% per weight) and bulk density (g/cm3) in the depth of 060 cm are shown in Table (2).


Table (2): Soil moisture constants of the experimental field at Giza Agricultural Research Station Depth (cm) 0 – 15 15 - 30 30 - 45 45 - 60

Field capacity (%, w/w) 41.85 33.68 28.36 28.05

Wilting point (%, water) 18.61 17.50 16.92 16.54

Available water (mm) 40.0 30.1 20.6 22.1

Bulk density g/cm3 1.15 1.24 1.20 1.28

Metrological data were collected for Giza Agricultural Research Station and are included in Table (3). Table (3): Meteorological data for Giza region in 2006/07 and 2007/08 growing seasons 2006/07 growing season WS RH SS SR Epan Month 2 (m/s) (%) (h) (cal/cm /day) (mm/day) November 3.6 67 8.2 326 2.5 December 3.0 69 7.0 268 2.0 January 3.4 70 7.0 280 2.0 February 3.4 62 7.9 453 3.4 March 4.4 59 8.6 441 4.2 April 5.2 27.8 9.6 519 5.3 2007/08 growing season Epan Tmax Tmin WS RH SS SR Month (ºC) (ºC) (m/s) (%) (h) (cal/cm2/day) (mm/day) November 26.8 15.7 3.6 62 8.2 326 3.2 December 22.7 11.2 3.0 66 7.0 268 2.0 January 18.0 7.2 3.4 62 7.0 280 2.2 February 20.6 8.1 3.4 53 7.9 453 3.3 March 27.4 13.1 4.4 47 8.6 441 3.5 April 30.4 15.7 5.2 44 9.6 519 5.7 Tmax=Maximum temperature; TMin=Minimum temperature; WS=Wind speed; RH=Relative humidity; SS=Actual sunshine duration; SR= Solar radiation; Epan=Evaporation pan. Tmax (ºC) 23.9 20.8 19.5 21.6 24.6 27.8

Tmin (ºC) 14.2 11.2 9.0 11.6 13.2 16.1

Consumptive water use was calculated using soil sampling. Consumptive water use was estimated by the following equation (Israelsen and Hansen, 1962): CWU = (

2

−

1)

* Bd * ERZ

(2)

Where: CWU=the amount of consumptive use (mm), 2=soil moisture percentage after irrigation, 1=soil moisture percentage before the following irrigation, Bd=bulk density in g/cm3, ERZ= effective root zone.


Maximum leaf area index was measured. Harvest was done in the 3rd week of April during the two growing seasons. Wheat grain and biological yield were measured and harvest index was determined. Water use efficiency (kg/m3) values for the three varieties were calculated by the following equation (Vites, 1965): WUE= Grain yield (kg/ha)/Consumptive use (m3/ha)

(3)

2. CropSyst model calibration and validation CropSyst (Cropping Systems Simulation Model) is a multi-year, multi-crop, daily time step crop growth simulation model, developed with emphasis on a friendly user interface, and with a link to GIS software and a weather generator (Stockle, 1994). The model’s objective is to serve as an analytical tool to study the effect of cropping systems management on crop productivity and the environment. For this purpose, CropSyst simulates the soil water budget, soil-plant nitrogen budget, crop phenology, crop canopy and root growth, biomass production, crop yield, residue production and decomposition, soil erosion by water, and pesticide fate. These are affected by weather, soil characteristics, crop characteristics, and cropping system management options including crop rotation, variety selection, irrigation, nitrogen fertilization, pesticide applications, soil and irrigation water salinity, tillage operations, and residue management. After each growing season, input files required by CropSyst model for Giza location and wheat crop were prepared and use to run the model. A few variety-specific parameters were calibrated within a reasonable range of fluctuation set in CropSyst manual. After calibration, the model was validated using the measured data of the three varieties for grain and biological yield and consumptive use. To test the goodness of fit between the measured and predicted data, percent difference between measured and predicted values for each variety in each growing season were calculated, in addition to root mean squared error (Jamieson, et al., 1998) and Willmott index of agreement (Willmott, 1981). Furthermore, regression analysis was done to test the strength of the relationship between measured and predicted yield and consumptive water use values.

3. Climate change scenarios In this work, the HadCM3 which is a coupled atmosphere-ocean general circulation model (AOGCM) developed at the Hadley Centre for Climate Prediction and Research (United Kingdom) was used (Gordon et al., 2000 and Pope et al., 2000) and considered as significantly and more sophisticated than earlier versions (Hulme et al., 1998). This model has a spatial resolution of 2.5 x 3.75 (latitude by longitude). HadCM3 provide information about climate change all over the entire world during the 21st century and present information about three times slices: 2020s, 2050s, and 2080s. In order to provide information on possible changes in the world climate, the climate change


models are forced to consider future scenarios. The IPCC (Nakicenvic et al., 2000) has developed emission scenarios known as SRES (Special Report on Emission Scenarios). The four SRES scenarios combined two sets of divergent tendencies: one set varying between strong economic values and strong environmental values, the other set between increasing globalization and increasing regionalization (IPCCTGCIA, 1999). Two climate change scenarios were considered in this study: A2 and B2. These selected two scenarios: A2 and B2 consider a rise in global annual mean temperature by 3.09 and 2.16°C, respectively, CO2 concentration 834 and 601 ppmv, respectively and global mean sea level rise 62 and 52 cm, respectively. As the resolution of the model is too big, using simple interpolation techniques of these percentages have been applied to fit the station site. Data were downloaded in GRIB format from the IPCC Data Distribution Centre web site, and the GRBCONV program source code is found at the following web site: [http://www/dkrz.de/ipcc/ddc/html/HadleyCM3/hadcm3. html]. The GRBCONV program was used to convert the data files from GRIB format to the more conventional ASCII. The download site does not offer the option to subset the data based on an area of interest, so a custom program was used to extract the data for the region of interest. HadCM3 variables were monthly precipitation, solar radiation, minimum and maximum temperatures. A2 and B2 climate change scenarios were used to run the CropSyst model to predict wheat yield and consumptive use in the year of 2038. The reason for choosing that year to predict potential wheat yield is to perceive how wheat productivity will be affected after 30 years. The effect of climate change on each of the two growing season will be discussed separately as if each season could be a representation of the growing season of the year of 2038.

4. Adaptation strategies The effect of two early sowing dates and irrigation rescheduling on wheat yield was tested under the two climate change scenarios was investigated using CropSyst model. The proposed sowing dates were planting in the 1st of November and on 21st of October. The proposed irrigation scenario suggested to increase the number of irrigation from 7 irrigations to 8 irrigations and to apply irrigation every 21 days to refill plant available water to prevent the occurrence of water stress. Furthermore, Table (4) showed actual irrigation schedule in the two growing seasons and proposed irrigation schedule.


Table (4): days after planting for each actual single irrigation for the two growing seasons and for the proposed irrigation schedule Irrigation Proposed irrigation Actual irrigation date number 2006/07 growing season 2007/08 growing season date st 1 Planting day Planting day Planting day nd 2 30 30 21 3rd 51 58 42 4th 75 77 63 th 5 98 94 84 6th 121 112 105 7th 142 134 126 th 8 ----147 Harvest 159 157 150-157

RESULTS 1. CropSyst model validation 1.1. Wheat grain yield prediction

Table (5) shows measured versus predicted wheat yield in the two growing seasons. Results in that table implied that CropSyst model predicted wheat yield with high degree of accuracy. Percent difference between measured and predicted wheat yield was less than 1%. RMSE was 0.0157 ton/ha and Willmott index of agreement was 0.9999. Table (5): Measured versus predicted wheat grain yield (ton/ha) in the two growing seasons 2006/07 growing season 2007/08 growing season Measured Predicted Percent Measured Predicted Percent yield yield reduction yield yield reduction Sids 1 5.92 5.91 0.20 5.40 5.39 0.19 Sakha 93 5.86 5.82 0.64 5.39 5.36 0.61 Giza 168 5.52 5.51 0.16 5.38 5.38 0 RMSE 0.0157 WI 0.9999 RMSE= root means square error; WI= Willmott index of agreement. Variety

Results in Figure (1) imply that all predicted wheat values lies within 95% confidence interval (95% CI). Regression analysis of the measured and predicted wheat yield values indicated a significant relationship (P